Software Engineering Courses (SECourses)•9mo ago

I always believed that VRAM was the same as RAM? This party always confuses me!

JJdotBdot I always believed that VRAM was the same as RAM? This party always confuses me!

Furkan Gözükara SECourses•4/8/25, 11:02 AM

totally different

Furkan Gözükara SECourses•4/8/25, 11:02 AM

RAM is system RAM and VRAM is GPU RAM

JJdotBdot Hello, it’s an eGPU RTX3080 with 16GB RAM, task manager says it’s all available....

Furkan Gözükara SECourses•4/8/25, 11:02 AM

you can show me screenshot of task manager

JJdotBdot I always believed that VRAM was the same as RAM? This party always confuses me!

404_error•4/9/25, 4:23 AM

Vram is a type of ram only for the gpu, it has way better bandwitch and optimized for certain type of data and lot of transfer. But it has a lot more latency that the ram and it too slow for random acces file and slow data that's why we got ram and vram

404_error•4/9/25, 4:25 AM

DDR5 in dual channel can transfer like 90GB/s, the GDDr7 of the 5090 can do 1792GB/s. Nothing compare

404_error•4/9/25, 2:19 PM

@Furkan Gözükara SECourses
Hello,

i run a dreambooth training on my machine and when i leave my home and come back i saw the training crash with no explication. I have a 10epoch checkpoint, (i think i was near 15), how i can resume the training ?

JdotBdotOP•4/9/25, 2:52 PM

Hey thanks for sharing the info on Vram , I feel that I understand it now but still not entirely sure how to measure what Vram I have! I’ve attached a screenshot from my task manager maybe someone can take a look, knowing what Vram I have will help me choose the correct Koyha Flux workflow, thanks in advance!

4404_error <@205854764540362752> Hello, i run a dreambooth training on my machine and when...

Furkan Gözükara SECourses•4/9/25, 3:47 PM

probably out of RAM

Furkan Gözükara SECourses•4/9/25, 3:47 PM

yes you can give last saved checkpoint

Furkan Gözükara SECourses•4/9/25, 3:47 PM

and it will continue - just reduce number of epochs you gonna do

JJdotBdot Hey thanks for sharing the info on Vram , I feel that I understand it now but st...

Furkan Gözükara SECourses•4/9/25, 3:47 PM

yes you have 32 GB RAM that is the issue

Furkan Gözükara SECourses•4/9/25, 3:47 PM

please upgrade to min 64 GB

Furkan Gözükara SECourses•4/9/25, 3:47 PM

more is better

JJdotBdot Hey thanks for sharing the info on Vram , I feel that I understand it now but st...

404_error•4/9/25, 8:16 PM

You have 16GB vram, it's write in "Dedicated GPU memory"

JdotBdotOP•4/10/25, 12:06 AM

Thanks for checking that for me! I’ll probably take the good Doc’s advice and upgrade soon — but in the meantime, out of the Lora training workflows available, which one would you recommend I try with my current setup?

JJdotBdot Thanks for checking that for me! I’ll probably take the good Doc’s advice and up...

Furkan Gözükara SECourses•4/10/25, 9:16 AM

sadly none will work very possibly but you can try this

Furkan Gözükara SECourses•4/10/25, 9:16 AM

use fp8 dev flux

Furkan Gözükara SECourses•4/10/25, 9:16 AM

use fp8 config

Furkan Gözükara SECourses•4/10/25, 9:16 AM

and select fp8 base on gradio interface and let me know if works

JdotBdotOP•4/11/25, 2:02 AM

Thanks Doc! Rather than struggle along I feel it will be more productive that I spend my time learning how to set up Massed Compute for Lora training for now, I’ll give it a whirl!

JJdotBdot Thanks Doc! Rather than struggle along I feel it will be more productive that I ...

Furkan Gözükara SECourses•4/11/25, 10:41 AM

great mindset

Original message was deleted

Furkan Gözükara SECourses•4/11/25, 2:51 PM

for SDXL? that can work yes

Furkan Gözükara SECourses•4/11/25, 2:51 PM

but for flux they will get mixed

Furkan Gözükara SECourses•4/11/25, 2:52 PM

ah you mean their defualt configs?

Furkan Gözükara SECourses•4/11/25, 2:52 PM

they never worked good for me